The Research Of Domain Term Extraction And Constructed Model Of Semantic Conceptual Graphs

With the development of natural language processing technology, semantic analysistechnology began to be applied to all areas of Chinese information processing. Semanticanalysis based on the conceptual graph is hot and research trends. However, the traditionalmethod of construction of semantic conceptual graphs which put syntactic analysis as the core,is not suitable to build Chinese semantic conceptual graphs. This is because the Chineselanguage which has characteristic of concept coupling, has not rich markup of syntax,semantic interact with syntax. Syntax analysis is difficult. Therefore, only combining with anumber of key technologies, can we build a breakthrough in terms of Chinese semanticconceptual graphs.This paper studies two key technologies of constructing Chinese semantic conceptualgraphs: techniques of term recognition and extraction, and the model of constructing sentencesemantic conceptual graphs. On the basis of original study, we propose a new method forautomatic term extraction, and establish a hierarchical model of constructing sentencesemantic conceptual graphs. The main researches are as follows:First, on the basis of existing studies, we propose a new method for automatic termextraction. It introduces the background corpus into C-value method, proposes the concepts ofword field distribution degree and effective word frequency, and improves the extractionperformance of terms combined with the term cluster recognition and mining. The termextraction experiment in the computer field shows that the proposed improved method(EC-value method) can measure the termhood of terms more effectively, and improve theextraction performance of low-frequency terms．Second, on the basis of existing research this paper proposes a assumption of conceptintegration and hierarchical recursion structure of E-A-V graph, and establish a hierarchicalmodel of constructing sentence semantic conceptual graphs. This model will make the tasksof constructing semantic conceptual graphs hierarchical and step by step. The model willdecompose a complex process into a few of basic tasks, and make us ensure the basic step ofconstructing semantic conceptual graphs. Experiment shows that the hierarchical model ofconstructing sentence semantic conceptual graphs is effective and can construct real semanticconceptual graph.